In the early 1970s, two laser physicists put a young Israeli showman in a shielded room and set out to prove he was a fraud. What they came away with was a paper in Nature, one of the most prestigious journals on Earth, reporting that under their controls Uri Geller had reproduced randomly chosen drawings he could not have seen. That a serious journal published it at all is the part the story usually skips, and it is the part that still bites.

Here is the setup, on the record. Russell Targ and Harold Puthoff at the Stanford Research Institute (SRI) ran Geller through a series of trials over several weeks in 1972 and 1973. In the central experiment, a target picture was selected, often by people who then stayed isolated from Geller, and Geller, in a separate shielded room, tried to draw it. Targ and Puthoff also reported a striking die-in-a-box test: a die was shaken inside a closed steel box and Geller called the upturned face correctly eight times out of eight attempts, declining to guess twice, odds they put around a million to one.

The documentary evidence is real and unusually accessible. The paper, "Information transmission under conditions of sensory shielding," appeared in Nature on 18 October 1974. Crucially, Nature's editors did something almost unheard of: they ran an accompanying unsigned editorial explaining why they were publishing such a weak and controversial paper, essentially as a specimen of what parapsychology research looked like, with the controls openly criticized. The drawing-replication trials are also preserved on film; SRI footage of Geller reproducing target sketches exists and can be watched. This is not folklore. It is a journal of record and a paper trail.

And then the controls fell apart under scrutiny, which is the skeptical reading and a sound one. The magician James Randi and the psychologist Ray Hyman dissected the protocols and found gaps a stage performer could drive a truck through: opportunities for Geller to glimpse targets, sensory leakage, loose isolation, and the simple fact that the experimenters were physicists, not conjurors, and so were not trained to catch the one thing Geller demonstrably was, a skilled sleight-of-hand artist. Even the Nature editorial flagged the experimental design as inadequate. Subsequent investigators, including people who worked around SRI, came to regard Geller's results as products of trickery and poor blinding rather than any transmitted information.

But foreground the inconvenient detail the debunk often glides past. The reason Geller mattered to SRI in the first place was that the work was entangled with government interest in psychic phenomena, the same SRI program that produced the remote-viewing research later folded into the intelligence community's STARGATE effort, whose files the CIA has since declassified. Sober, security-cleared people spent real money trying to determine whether any of this was operationally usable. The Nature paper was not a tabloid stunt; it was the public face of a research line that serious institutions funded for two decades.

The fair conclusion is uncomfortable from both ends. The evidence does not support the claim that Geller has paranormal powers; the controls were too weak to exclude cheating, and a trained magician should have been in the room. Equally, the historical fact stands that trained physicists, writing in Nature, could not in the moment explain how a performer produced their results, and admitted as much in print. That gap between "we cannot explain it" and "it is real" is exactly where pseudoscience lives, and exactly where Geller built a career.

So the unresolved question is not whether spoons bend by mind, the evidence says no. It is sharper and more about us: how did a sleight-of-hand act get through peer review at the most respected journal in science, and what does it say that the editors knew the controls were bad and printed it anyway? The spoon was never the mystery. The institutions that took the spoon seriously are.